ThesisPDF Available

Heart rate variability is representative of individual training adaptation to an altitude training camp in elite triathletes

Authors:
  • HRV4Training

Abstract and Figures

1 Summary Objectives: To determine if changes in resting heart rate (HR) and heart rate variability (HRV) during the beginning (first 10 days) of a three-week training camp at altitude are representative of the athlete's training adaptation at the end of the training camp. Methods: Four elite triathletes (2 male and 2 female), spent 23 days in Namibia at 1655 meters of altitude for two consecutive years in January 2019 and January 2020. Resting HR and HRV (the root mean square of successive RR intervals, or rMSSD) were measured daily upon wakening, while training data (GPS, heart rate) was acquired during cycling and running workouts. The athlete group was divided in responders and non-responders to the training camp at altitude based on the ratio between velocity and heart rate during aerobic running workouts of moderate duration. In particular, athletes whose velocity to heart rate ratio during workouts at week three was within the athlete's smallest worthwhile change (SWC) of pre-camp values, were considered responders. The difference in resting heart rate, rMSSD and the coefficient of variation of rMSSD (CV rMSSD) between the two weeks prior to the training camp and the beginning (first 10 days) of the training camp were computed as potential markers of future training adaptation. Results: Resting HR was significantly more elevated during the beginning of the training camp for non-responders (N = 3, HR difference = +4.6 bpm) with respect to responders (N = 4, HR difference = +0.5 bpm, p = 0.023). The CV rMSSD also increased by a greater extent for non-responders (+10%) with respect to re-sponders (-3%, p = 0.015). The difference in rMSSD was lower during the first week of camp for non-responders (-10 ms) with respect to responders (+6 ms), but this difference was not significant (p = 0.336). Conclusions: Athletes that responded positively to a three-week training camp at altitude showed a more favorable physiological response during the beginning of the training camp (smaller resting HR difference, lower CV rMSSD). This information can be used to further adjust training plans at the individual level.
Content may be subject to copyright.
Human Movement Sciences
Sport, Exercise & Health
Research Project 2019-2020
Heart rate variability is representative of
individual training adaptation to an altitude
training camp in elite triathletes
Vrije Universiteit Amsterdam
Faculty of Behavioural and Movement Sciences
Department of Human Movement Sciences
Qualification: MSc in Human Movement Sciences
Research Project
Author: Marco Altini
Student number: 2659875
Supervisors/examiners
Supervisor/examiner 1: Prof.dr. Thomas Janssen
Daily supervisor: Drs. Sander Berk (Appl. Scientist Dutch Triathlon Fed.)
Second examiner: Dr. Jos de Koning
June 2020
RM research report
1
Summary
Objectives: To determine if changes in resting heart rate (HR) and heart rate variability
(HRV) during the beginning (first 10 days) of a three-week training camp at
altitude are representative of the athlete’s training adaptation at the end of the
training camp.
Methods: Four elite triathletes (2 male and 2 female), spent 23 days in Namibia at 1655
meters of altitude for two consecutive years in January 2019 and January 2020.
Resting HR and HRV (the root mean square of successive RR intervals, or
rMSSD) were measured daily upon wakening, while training data (GPS, heart
rate) was acquired during cycling and running workouts. The athlete group was
divided in responders and non-responders to the training camp at altitude
based on the ratio between velocity and heart rate during aerobic running
workouts of moderate duration. In particular, athletes whose velocity to heart
rate ratio during workouts at week three was within the athlete’s smallest
worthwhile change (SWC) of pre-camp values, were considered responders.
The difference in resting heart rate, rMSSD and the coefficient of variation of
rMSSD (CV rMSSD) between the two weeks prior to the training camp and the
beginning (first 10 days) of the training camp were computed as potential
markers of future training adaptation.
Results: Resting HR was significantly more elevated during the beginning of the training
camp for non-responders (N = 3, HR difference = +4.6 bpm) with respect to
responders (N = 4, HR difference = +0.5 bpm, p = 0.023). The CV rMSSD also
increased by a greater extent for non-responders (+10%) with respect to re-
sponders (-3%, p = 0.015). The difference in rMSSD was lower during the first
week of camp for non-responders (-10 ms) with respect to responders (+6 ms),
but this difference was not significant (p = 0.336).
Conclusions: Athletes that responded positively to a three-week training camp at altitude
showed a more favorable physiological response during the beginning of the
training camp (smaller resting HR difference, lower CV rMSSD). This infor-
mation can be used to further adjust training plans at the individual level.
Key words
Heart rate variability; training adaptation; triathlon; altitude training; resting heart rate;
RM research report
2
Table of Contents
1 Introduction .............................................................................................. 3
2 Methods & Procedures ................................................................................. 4
2.1 Subjects ................................................................................................ 4
2.2 Procedures and data acquisition ................................................................ 4
2.3 Data analysis .......................................................................................... 4
2.4 Statistics ................................................................................................ 7
3 Results ..................................................................................................... 8
4 Discussion ................................................................................................. 8
4.1 Limitations ............................................................................................ 10
4.2 Recommendations for practitioners and researchers .................................... 11
5 Conclusions .............................................................................................. 11
6 Acknowledgements .................................................................................... 11
7 References ............................................................................................... 12
RM research report
3
1 Introduction
Resting heart rate (HR) and heart rate variability (HRV) have long been used to monitor
athletes recovery from previous workouts [1, 2]. The rationale behind monitoring recovery
using resting HR or HRV is that heavy training shifts the autonomic nervous system towards
a sympathetic drive [3], which is reflected in higher HR and lower HRV within 24h to 48h
after training. Among HRV features, the square root of the mean squared difference be-
tween beat to beat intervals (rMSSD) has emerged as a clear marker parasympathetic
activity. Additionally, the coefficient of variation of rMSSD (CV rMSSD) is often used as a
marker of day to day variability in HRV and representative of training adaptation [4]. Re-
ductions in rMSSD and increases in HR as measured at rest first thing in the morning, the
day after high intensity aerobic exercise, have been reported across a wide range of indi-
viduals [5, 6]. In recent times, monitoring HR and HRV unobtrusively in real-life settings,
outside of the lab, has finally become a practical possibility. Data can be acquired using
validated smartphone apps [7, 8] longitudinally over periods of weeks or months, providing
novel insights on an athlete’s response to training and lifestyle stressors. As a result, mon-
itoring recovery status and training adaptation by means of an HRV measurement is be-
coming more common among elite athletes as well as sports enthusiasts [9].
In elite sports, it is of great interest to understand if an athlete is adapting well to a training
program (e.g. a certain type of periodization). Yet, proper interpretation of HRV trends in
the medium and long term (weeks to months) is challenging as individuals can respond
differently to the training stimulus, and additional stressors (e.g. lifestyle) can play a role.
Insights from the application of HRV analysis in elite athletes is scarce as most research
has been carried out in college students and for limited periods of time. Additionally, most
studies focus on the relation between HRV and training load, while it would be of greater
practical interest for coaches and applied scientists, to determine how training load and
the subsequent physiological response (e.g. change in HRV) impact training adaptation or
performance. Finally, the specific demands of a given sport and environment could further
impact changes in physiology. For example, in triathlon, coaches often organize training
camps that aim at significantly increasing training demands so that positive adaptations
are triggered, and very limited data is available on the relationship between resting phys-
iology (HR, HRV) and training adaptation during a training camp with given characteristics.
Training camps provide a unique opportunity for coaches to manipulate training-related
variables (duration and intensity of the sessions for example) in an attempt to trigger
specific adaptations that should result in improved performance. Similarly, environmental
variables could be manipulated while maintaining training variables constant, for example
during a training camp at altitude or in the heat. While the rationale behind training at
altitude is clear and well documented, with an increase in red blood cells and improved
oxygen carrying capacity resulting from spending time in a hypoxic environment, athletes
often respond differently and not always positively. Some athletes might adapt to the novel
stimulus and improve, while others might struggle or require more time to adapt [10, 11].
In this context, HRV could provide a simple, non-invasive assessment of the ability of each
athlete to deal with the stressors imposed by the training camp (e.g. increased training
load, traveling, spending time at altitude, etc.) and could allow the coach to further adjust
training plans at the individual level. Based on published literature, a favorable profile in
terms of resting physiology would mean a smaller or absent increase in resting HR, reduc-
tion in rMSSD and increase in CV rMSSD [12, 4]. The ability to detect, and potentially
identify in advance, multiparameter physiological trends representative of positive and
negative adaptation can be key in training individualization and optimization during or after
the camp. For example, reducing training load for individual athletes who demonstrate
unfavorable physiological responses (e.g. reduced HRV and increased CV) in response to
intense training may help prevent excess fatigue accumulation. Even of more practical
interest would be to determine if changes in resting HR and HRV during the beginning of
the training camp are representative of training adaptation at the end of the camp, so that
RM research report
4
the coach can potentially make adjustments early during the camp, and not only use HR
and HRV as a method of assessment of adaptation once the camp is over.
Thus, our research question is: are changes in resting HR and HRV during the beginning
of a 3-week training camp at altitude representative of the athlete’s training adaptation at
the end of the training camp?
2 Methods & Procedures
2.1 Subjects
Four elite triathletes (2 male and 2 female), spent 23 days in Namibia at 1655 meters of
altitude for two consecutive years in January 2019 and January 2020, providing a total of
N = 7 individual responses to the training camp at altitude. Data from one athlete during
the training camp in 2019 did not include measures of resting physiology and could not be
used for the analysis. All triathletes provided written consent form after being informed
about the study and analysis to be carried out. The study was approved by The Scientific
and Ethical Review Board (VCWE) of the Faculty of Behavior & Movement Sciences, Vrije
Universiteit Amsterdam.
2.2 Procedures and data acquisition
HR and HRV measurements at rest were taken using the validated HRV4Training app [8].
The HRV4Training app is a commercially available mobile phone app [7] which allows for
non-invasive measurements of HR and HRV using either the phone camera, or an external
sensor. The triathletes were instructed to take measurements first thing in the morning,
before other confounding factors can have an impact (coffee, physical activity, etc.), ideally
while lying down in bed. Athletes were also instructed to measure daily. Timestamped
heart rate and rMSSD data were downloaded from the HRV4Training Pro web platform to
a csv file for further processing. Workouts data were collected using GPS watches paired
to heart rate monitors (for running workouts) and heart rate monitors as well as power
meters (for cycling workouts). Workouts data were automatically exported to the Train-
ingPeaks platform, which is the training planning and analysis tool of choice of the triath-
letes coaching staff. Workouts summaries were then manually exported from Train-
ingPeaks to csv files for further processing.
2.3 Data analysis
The analysis does not include any data collected in laboratory settings but relies on meas-
urements and workouts collected in unsupervised free-living settings. Thus, data cleaning
is paramount. Similarly, different choices were made to determine training adaptation and
the ideal windows on which to analyze the data, which are motivated in the sections below.
2.3.1 Data cleaning
In terms of measurements of resting physiology, we have removed all measurements
whose quality was lower than 90% as reported by the HRV4Training app. This threshold
was determined empirically based on prior work with the measurements collected with the
app. Additionally, beat to beat intervals (RR intervals) and plethysmography (PPG) data
were visually analyzed by the experimenter, to determine if any artifacts were still present
after the artifact removal algorithm embedded in the app had already performed a step of
data cleaning. Workouts data were processed to cluster athletes between responders and
non-responders. In particular, we used running workout for this purpose, since running
workouts involve less variables and can be simpler to analyze with respect to cycling
workouts, were factors such as drafting, workout duration, etc. can confound the analysis.
Running data was used to determine the velocity to heart rate ratio, a marker of aerobic
RM research report
5
endurance, but also in this case of training adaptation, since heart rate at a given intensity
is expected to raise for an athlete during the initial phase of the camp at altitude, and to
subsequently reduce as a sign of positive adaptation. Running data was pre-processed as
follows: all workouts without velocity and heart rate data were excluded. Very short
workouts in which an athlete might have not reached a steady state in terms of heart rate
were excluded (workouts shorter than 25 minutes). Workouts with low average heart rates,
typically caused by sessions involving short intervals and long breaks, and therefore not
representative of aerobic capacity, were also excluded. Finally, given the potential issues
due to sensor malfunction and simply the fact that anything can happen outside of the lab,
we used an additional filter that removed workouts outside of the 15th and 85th percentiles,
so that outliers in terms of the relation between velocity and heart rate would be excluded.
2.3.2 Defining adaptation to the training camp
Athletes were divided in responders and non-responders to the training camp at altitude.
In particular, prior research has shown how training adaptation can be identified using
submaximal heart rate data, where a typical pattern of positive adaptation shows submax-
imal heart rate initially higher for a given intensity [13], and gradually reducing to sea level
values [14]. On the other hand, some athletes do not adapt and show lactate or submax-
imal heart rate levels similar to the ones of the first week even at the end of the training
camp, once again raising the important point of training individualization [14]. Thus, in our
study we considered responders the athletes that were able to positively adapt within the
duration of the camp.
Figure 1. An example of submaximal heart rate data (the velocity to heart rate ratio) grouped by
week and color-coded according to the weekly average being within or outside of the SWC. In this
example we can see how the first week of the training camp shows a large change in the velocity to
heart rate ratio, which reduces due to increased heart rate, as typical of altitude training. During the
second week, the ratio improves, but is still outside of the SWC, as shown by the darker gray color.
Finally, this athlete’s velocity to heart rate ratio during workouts is back within the SWC by week 3,
the last week of the training camp. Thus, the athlete is assigned to the responders group, or the
ones that have adapted.
0.023
0.024
0.025
0.026
0.027
Week before Week 1 Week 2 Week 3
Camp week
Speed over heart rate
Outside SWC
FALSE
TRUE
Velocity over heart rate for all running workouts, grouped by camp week
RM research report
6
In particular, we used the concept of smallest worthwhile change (SWC), computed as
30% of the standard deviation in the ratio between velocity and heart rate, for an individual
athlete, to determine which athlete is a responder. This means that if the velocity to heart
rate ratio is within the SWC during the last week of the training camp, the athlete is con-
sidered a responder. On the other hand, if the ratio of velocity to heart rate is still outside
of the SWC by the last week of the camp, then the athlete is considered a non-responder,
as heart rate is still suppressed to a greater degree than normal day to day variability
would allow. An example of submaximal heart rate data (the velocity to heart rate ratio)
grouped by week and color-coded according to the weekly average being within or outside
of the SWC, is shown in Figure 1.
Figure 2. Example of daily resting HR and HRV for one athlete, as well as boxplots of the same
parameters during the two periods of interest: before the camp and during the beginning of the
camp. We can see how resting heart rate increases during the first days of the camp, and then lowers
to values similar to the athlete’s values before the camp. Similarly, rMSSD reduces during the first
days of the camp, even though it normalizes quickly for this athlete, as shown by the boxplots which
highlight how there is no difference between the two periods. We can see however how rMSSD
spreads over a broader range, and jumps between higher and lower values, as also shown by a wider
boxplot. This behavior is what is captured by a higher CV, or higher day to day variability despite a
similar average rMSSD over a period of ten days.
40
50
60
70
Date
Heart rate (bpm)
Phase
Week before
First week
Other
Daily heart rate, colorcoded by training camp phase
50
75
100
125
150
Date
rMSSD (ms)
Phase
Week before
First week
Other
Daily rMSSD, colorcoded by training camp phase
40
50
60
70
Week before First week
Phase
Heart rate (bpm)
Phase
Week before
First week
Heart rate, grouped by training camp phase
50
75
100
125
150
Week before First week
Phase
rMSSD (ms)
Phase
Week before
First week
rMSSD, grouped by training camp phase
RM research report
7
2.3.3 Choice of time windows for HRV analysis
HRV measurements were collected almost daily for more than two years for three of the
elite triathletes taking part in this study. On the other hand, the fourth athlete had less
consistent measurements but did measure consistently during the second training camp,
including in the week leading to the camp. Previous research has shown how 5 measure-
ments over a period of a week are required to establish an accurate HRV baseline, given
the relatively high day to day variability in resting HRV [15]. Despite the well-known phys-
iological changes occurring when starting a training camp at altitude, e.g. higher resting
and workout heart rates [16], we hypothesized that responders would show a more favor-
able profile in terms of resting physiology. The difference in resting heart rate, rMSSD and
the coefficient of variation of rMSSD (CV rMSSD) between the week prior to the training
camp and the beginning of the training camp were computed as potential markers of future
training adaptation. In particular, we hypothesized that responders could show a lower
increase in resting heart rate and CV rMSSD or lower reduction in rMSSD with respect to
the week prior to the camp. On the other hand, non-responders could show a less favorable
resting physiology profile (e.g. resting heart rate increasing to a greater extent between
the week prior to the camp and the first week of the camp). The CV rMSSD, quantifies day
to day variability, and therefore benefits from additional data points. Thus, to allow for 1)
an assessment of an athlete’s normal physiology prior to the camp, 2) missing measure-
ments or poor quality measurements that had to be discarded, we used the two weeks
prior to the camp and the first ten days of the camp to compute differences in resting HR
and HRV. Figure 2 shows an example of daily resting HR and HRV for one athlete, as well
as boxplots of the same parameters during the two periods of interest: before the camp
and during the beginning of the camp.
Figure 3. An example of submaximal heart rate data (the velocity to heart rate ratio) for two athletes,
grouped by training camp phase and color-coded according to the weekly average being within or
outside of the SWC. The athlete on the left is part of the responders group, as the velocity to heart
rate ratio is within the SWC during the last week of the camp. On the other hand, the athlete on the
right is assigned to the non-responders group, as the velocity to heart rate ratio remains suppressed
even at the end of the camp.
2.4 Statistics
Differences between the time window before the camp (14 days) and the beginning of the
training camp (10 days) were computed for rMSSD, heart rate and CV rMSSD. Thus, each
athlete was characterized by her or his individual response to the training camp. Individual
responses (N = 7) were grouped into two groups, one for non-responders (N = 3, athletes
that did not adapt during the training camp, defined as the velocity to heart rate ratio being
0.023
0.024
0.025
0.026
0.027
0.028
Week before First week Last week
Phase
Speed over heart rate
Outside SWC
FALSE
TRUE
Velocity over heart rate for all running workouts, grouped by phase
0.020
0.021
0.022
Week before First week Last week
Phase
Speed over heart rate
Outside SWC
FALSE
TRUE
Velocity over heart rate for all running workouts, grouped by phase
RM research report
8
within the athlete’s SWC during the last week of the training camp) and one for responders
(N = 4). We used t-tests to determine if the difference in resting physiology (HR, rMSSD
and CV rMSSD) between the period prior to the camp and the beginning of the camp was
significantly different between responders and non-responders. Figure 3 shows an example
for two athletes, a responder and a non-responder. Finally, external training load was not
manipulated during the camp, as the load before and during the camp was very similar
(see Figure 4). To verify that training load is not a confounding factor in our analysis, we
also performed a t-test of the difference in training load between pre-camp and the begin-
ning of the camp in the two groups.
Figure 4. Example of daily external training load data (TSS) for one athlete, as well as boxplots of
the same parameters during the two periods of interest: before the camp and during the beginning
of the camp. We can see how training load is similar between the two phases, and most likely is not
the main factor behind changes in resting physiology analyzed in this study, which are instead asso-
ciated with responses to the altitude stimulus.
3 Results
Resting HR was significantly more elevated during the first week of the training camp for
non-responders (+4.6 bpm) with respect to responders (+0.5 bpm, p = 0.023). The CV
rMSSD also increased by a greater extent for non-responders (+10%) with respect to re-
sponders (-3%, p = 0.015). The difference in rMSSD was lower during the first week of
camp for non-responders (-10 ms) with respect to responders (+6 ms), but this difference
was not significant (p = 0.336). Figure 5 shows boxplots of the differences in the two
groups. As expected, training load did not differ before and during the camp and did not
differ between the two groups (responders and non-responders, p = 0.966).
4 Discussion
The main finding of this study is that athletes that responded positively to a three-week
training camp at altitude showed a more favorable physiological response during the be-
ginning of the training camp. In particular, responders showed a lower increase in resting
HR when going from sea level to altitude, with respect to non-responders. Additionally,
responders showed a lower increase in HRV and no increase in the CV rMSSD when going
from sea level to altitude. On the other hand, non-responders showed a larger increase in
resting HR when going from seal level to altitude, as well as a larger reduction in rMSSD
and larger increase in the CV rMSSD. To the best of our knowledge, this is the first time
that HRV profiles of elite triathletes during a training camp at altitude were analyzed with
respect to training adaptation at the end of the camp.
0
100
200
Date
TSS
Phase
Week before
First week
Daily workload, colorcoded by training camp phase
50
100
150
Week before First week
Phase
TSS
Phase
Week before
First week
Daily workload, grouped by training camp phase
RM research report
9
Figure 5. Boxplots of the difference in resting HR, rMSSD and CV rMSSD between the period prior to
the camp and the beginning of the camp, in the two groups (responders and non-responders). We
can see how responders have a more favorable physiological profile (lower increase in resting HR,
lower CV, less reduction in rMSSD, the latter being not significant according to the t-test).
It is clear both to the experienced coach and the scientist, that athletes respond differently
to the stimuli they are provided, and not all benefit in the same way from a given stimulus.
The basic principle behind training camps at altitude is to trigger an increase in red blood
cells, and therefore improve endurance capacity as an athlete develops an ability to carry
more oxygen. However, responses to training at altitude vary, as some athletes might
adapt to the novel stimulus and improve, while others might struggle or require more time
to adapt [10, 11]. Given the obvious interindividual differences, and the effectiveness of
measurements of resting physiology in capturing training responses and training adapta-
tion, our hypothesis was that responders to altitude training would show a more favorable
profile in terms of resting physiology. For example, responders would show a lower in-
crease in resting heart rate and CV rMSSD or lower reduction in rMSSD with respect to the
week prior to the camp.
A favorable physiological profile has been shown in previous research in different but some-
what similar settings. For example, the HRV of climbers during an ascent was predictive of
which climbers would develop altitude sickness in a following stage of the ascent [17].
Looking at training camps, physiological responses when transitioning between the off-
season and the preparatory phase of the season have been investigated several times in
team sports, showing consistent results. Typically, in these studies, altitude is not manip-
ulated, and the main stressor is the large increase in training load, coming from the off-
season. In response to an increase in training load with the beginning of the training camp,
the authors reported that the CV was the most sensitive marker, with larger CV being
reported during weeks of high training load. On the other hand, the mean rMSSD did not
change much [12]. This is in line with our finding in the current study, where the CV was
more sensitive to training adaptation than the average rMSSD. We speculate that a better
response to the training camp is due to less physiological stress being present (or perceived
by the athlete), which facilitates a more stable HRV profile (i.e. a lower CV). Flatt argues
that a lower CV possibly represents reduced perturbation in homeostasis and reducing
training load for individuals who demonstrate an unfavorable psychometric and physiolog-
ical profile in response to intense training, may help prevent excess fatigue accumulation.
In our study, we could make similar assumptions. In particular, the disruption in homeo-
stasis due to altitude, is evident in all athletes during the first days of the training camp,
both in resting and workouts data. Typical responses the first day show elevated heart rate
while running and elevated heart rate at rest for example. However, it seems that the
0
2
4
6
FALSE TRUE
Adapted
bpm
Heart rate, p = 0.023
20
0
20
40
FALSE TRUE
Adapted
ms
rMSSD, p = 0.336
0.05
0.00
0.05
0.10
FALSE TRUE
Adapted
%
Coefficient of variation, p = 0.015
RM research report
10
degree of the disruption in homeostasis is greater for certain athletes, and that these ath-
letes are the ones that eventually will not adapt to the training camp. In another study,
the same authors reported a large relationship between performance tests and changes in
the CV, indicating how the players that showed a decrease in the CV during the first 2
weeks of the training camp, experienced greater performance improvements post-training
camp [4]. These findings are in line with what was reported in endurance athletes, where
fatigued elite athletes had lower HRV and higher interindividual variation in HRV (CV),
compared to athletes that were not fatigued [18].
Proper interpretation of HRV trends in the medium and long term (weeks to months) is
more challenging than analysis of acute responses (e.g. the typical suppression in para-
sympathetic activity in the 24-48 hours following a hard workout). While the use of HRV
to monitor recovery and training adaptation is motivated by the acute drop we expect
following an intense workout or stressor, in the long term we also expect an increase in
HRV or reduction in day to day variability (CV rMSSD) following positive adaptation and
improved performance [19, 20]. This is not in contrast with the reduced HRV at the acute
level and reflects possible improved aerobic capacity or self-regulation in the long term
due to a positive adaptation to training. In particular, the few studies quantifying not only
HRV and external load, but also measuring performance, often showed stable or increased
HRV in response to high training load blocks for athletes that perform better. Typically,
this result is interpreted as the ability of the athlete to assimilate and positively adapt to
the training stimulus, as shown by increased performance. On the other hand, athletes
that do not improve performance, often show a poor physiological response, such as a
reduced HRV baseline or higher CV rMSSD [12, 21]. This also explains why it is incorrect
to correlate HRV and training load in the long term [22], as positive adaptation to training
should result in stable or increasing HRV.
In our study, external training load was similar before and during the camp, and the main
stressor was altitude. Yet, our findings are similar to published literature in terms of the
physiological profile of responders and non-responders, highlighting how even between
different sports and scenarios, resting HR, HRV and the CV are useful markers of adaptation
to a stressor. Thus, we speculate that similar considerations could apply, and manipulating
training load for athletes showing an unfavorable physiological response at the beginning
of the camp, might improve outcomes by the end of the camp.
4.1 Limitations
A limitation of this study is the small sample size. Unfortunately working with a small
sample size is typical of elite settings where, by definition, only a few athletes make it to
the top. Additionally, our analysis relies on 7 data points collected from 4 athletes but
treats the data points independently. Given the long time between camps (1 year), and
the fact that individual responses over time can always differ based on for example the
novelty, volume and intensity of the training stimulus, as well as the presence of others
forms of stress unrelated to training (e.g. lifestyle), we believe this is a valid approach.
Another potential limitation of the study is the use of real-life workouts data instead of
laboratory tests to determine adaptation to the training camp according to submaximal
data and the velocity to heart rate ratio. However, using data collected in unconstrained
free-living settings has several advantages. First, no specific tests are required, which
might cause additional anxiety, and a disruption in training, as well as the less than ideal
scenario in which a training day is used to acquire a data point via a standardized test
instead of prescribing a workout from which the athlete could benefit more. Secondly,
individual data points might be affected by various parameters, starting with the day to
day variability in heart rate, and using workouts data to determine training adaptation by
pooling several days of data, might be a more robust strategy, once data has been properly
filtered to remove outliers, as reported in this work. Finally, our analysis shows results
consistent with laboratory tests, where a typical pattern of positive adaptation to altitude
RM research report
11
training involves submaximal heart rate initially higher for a given intensity [13], and grad-
ually reducing to sea level values [14, 23]. The superiority of free-living data is especially
true for measurements of resting HR and HRV, where daily measurements taken first thing
in the morning, provide a better picture than isolated laboratory based assessments, in a
variety of applications [24].
4.2 Recommendations for practitioners and researchers
Monitoring resting physiology during a training camp at altitude can provide early insights
on which athletes will respond positively to the training camp, and which athletes will not
respond positively or adapt. Practitioners could use morning measurements of resting heart
rate and HRV during the first week of a training camp, and in particular assess the degree
of the increase in resting heart rate or in the CV rMSSD as signs that are potentially pre-
dictive of positive or negative adaptation later on. Once the athletes with this specific pro-
file have been identified, various strategies could be employed, either in terms of training
load manipulation or by modifying other parameters. Some examples could be having the
athletes stay at altitude for longer to determine if they require more time to adapt to the
stimulus, or manipulate more variables at the same time, for example having the athlete
sleep at altitude while training at sea level (sleep high train low). Interestingly, in this
study the same athletes responded in the same way on the two consecutive camps one
year apart, with two athletes always adapting and two athletes not adapting (data not
shown). Given the consistent response at the individual level for each athlete, it might be
of interest to implement some of these changes in the non-responders.
In terms of recommendation for future research, it would be helpful to replicate the findings
of this study on a larger population as well as on athletes of different level. Similarly, the
same analysis could be carried out at different altitudes, or in response to different stress-
ors, for example the heat, a major concern for the current Olympic cycle as well as long
course triathletes racing in Kona. It should also be noted how in this study we assessed
only one aspect (the change in submaximal heart rate while running) and performance is
a multifaceted construct, especially in a multi-sport event such as triathlon. Athletes that
were assigned to the non-responders group might have still improved via other pathways,
and it would be of interest to determine if a more favorable physiological response such as
the one documented in this study is not only representative of training adaptation, but also
of performance improvements. Carrying out such a study involves difficulties associated
with the small number of athletes typical of elite settings, as well as with the different
competitions the athletes normally take part in, which might make it challenging to deter-
mine performance gains in a standardized manner.
5 Conclusions
In this work, we have shown how athletes that responded positively to a three-week train-
ing camp at altitude had a more favorable physiological profile during the beginning of the
training camp (lower resting HR difference, lower CV rMSSD, higher rMSSD difference). To
the best of our knowledge, this is the first time that HRV responses of elite triathletes
during a training camp at altitude were analyzed with respect to training adaptation at the
end of the camp. Our findings open the door to individualized monitoring and just in time
interventions, as athletes whose physiological profiles show suboptimal results at the be-
ginning of the camp, could undergo training plan adjustments in the remaining of the camp.
6 Acknowledgements
The author would like to thank the athletes and coaches that have provided invaluable
resources in terms of data, time and knowledge. No funding was received for this study.
RM research report
12
7 References
[1]
Jeukendrup, "Physiological changes in male competitive cyclists after two weeks of intensified
training," International journal of sports medicine, pp. 534--541, 1992.
[2]
Plews, "Training adaptation and heart rate variability in elite endurance athletes: opening the
door to effective monitoring," Sports medicine, pp. 773--781, 2013.
[3]
Stanley, "Cardiac parasympathetic reactivation following exercise: implications for training
prescription," Sports medicine, pp. 1259--1277, 2013.
[4]
Flatt, "Evaluating individual training adaptation with smartphone-derived heart rate variability
in a collegiate female soccer team," The Journal of Strength & Conditioning Research, pp. 378-
385, 2016.
[5]
Dressendorfer, "Increased morning heart rate in runners: a valid sign of overtraining?," The
Physician and sportsmedicine, pp. 77--86, 1985.
[6]
Altini, "Hrv4training: Large-scale longitudinal training load analysis in unconstrained free-living
settings using a smartphone application," in 2016 38th Annual International Conference of the
IEEE Engineering in Medicine and Biology Society (EMBC), 2016.
[7]
Altini, "HRV4Training," 2013. [Online]. Available: https://www.hrv4training.com/.
[8]
Plews, "Comparison of heart-rate-variability recording with smartphone
photoplethysmography, Polar H7 chest strap, and electrocardiography," International journal
of sports physiology and performance, pp. 1324-1328, 2017.
[9]
Pichot, "Relation between heart rate variability and training load in middle-distance runners,"
Medicine & Science in Sports & Exercise, pp. 1729--1736, 2000.
[10]
Fulco, "Maximal and submaximal exercise performance at altitude," Aviation Space and
Environmental Medicine, pp. 793-801, 1998.
[11]
Friedmann-Bette, "Classical altitude training," Scandinavian journal of medicine & science in
sports, pp. 11-20, 2008.
[12]
Flatt, "Smartphone-derived heart-rate variability and training load in a women’s soccer team,"
International journal of sports physiology and performance, pp. 994--1000, 2015.
[13]
Taralov, "Heart rate variability as a method for assessment of the autonomic nervous system
and the adaptations to different physiological and pathological conditions," Folia medica, pp.
173-180, 2016.
[14]
Rusko, "New aspects of altitude training," The American journal of sports medicine, pp. S48-
S52, 1996.
[15]
Plews, "Monitoring training with heart-rate variability: How much compliance is needed for
valid assessment?," International journal of sports physiology and performance, pp. 783-790,
2014.
[16]
Hughson, "Sympathetic and parasympathetic indicators of heart rate control at altitude studied
by spectral analysis," Journal of Applied Physiology, pp. 2537-2542, 1994.
[17]
Karinen, "Heart rate variability changes at 2400 m altitude predicts acute mountain sickness
on further ascent at 3000-4300 m altitudes," Frontiers in physiology, p. 336, 2012.
[18]
Schmitt, "Fatigue shifts and scatters heart rate variability in elite endurance athletes," PloS
one, 2013.
[19]
Nakamura, "Ultra-short-term heart rate variability is sensitive to training effects in team sports
players," Journal of sports science & medicine, p. 602, 2015.
[20]
Vesterinen, "Predictors of individual adaptation to high-volume or high-intensity endurance
training in recreational endurance runners," Scandinavian journal of medicine & science in
sports, pp. 885-893, 2016.
[21]
Flatt, "Interpreting daily heart rate variability changes in collegiate female soccer players," J
Sports Med Phys Fitness, pp. 907--15, 2017.
[22]
Saw, "Monitoring the athlete training response: subjective self-reported measures trump
commonly used objective measures: a systematic review," Br J Sports Med, vol. 50, no. 5, pp.
281-291, 2016.
[23]
Ventura, "Intermittent hypobaric hypoxia induces altitude acclimation and improves the lactate
threshold," Aviation, space, and environmental medicine, pp. 125-30, 2000.
RM research report
13
[24]
Kokts-Porietis, "The effect of the menstrual cycle on daily measures of heart rate variability in
athletic women," Journal of Psychophysiology, 2019.
ResearchGate has not been able to resolve any citations for this publication.
Article
Full-text available
Purpose: To establish the validity of smartphone photoplethysmography (PPG) and heart rate sensor in the measurement of heart rate variability (HRV). Methods: 29 healthy subjects were measured at rest during 5 min of guided breathing (GB) and normal breathing (NB) using Smartphone PPG, heart rate chest strap and electrocardiography (ECG). The root mean sum of the squared differences between R-R intervals (rMSSD) was determined from each device. Results: Compared to ECG, the technical error of estimate (TEE) was acceptable for all conditions (average TEE CV% (90% CI) = 6.35 (5.13; 8.5)). When assessed as a standardised difference, all differences were deemed "Trivial" (average std. diff (90% CI) = 0.10 (0.08; 0.13). Both PPG and HR sensor derived measures had almost perfect correlations with ECG (R = 1.00 (0.99; 1:00). Conclusion: Both PPG and heart rate sensor provide an acceptable agreement for the measurement of rMSSD when compared with ECG. Smartphone PPG technology may be a preferred method of HRV data collection for athletes due to its practicality and ease of use in the field.
Article
Full-text available
The autonomic nervous system controls the smooth muscles of the internal organs, the cardiovascular system and the secretory function of the glands and plays a major role in the processes of adaptation. Heart rate variability is a non-invasive and easily applicable method for the assessment of its activity. The following review describes the origin, parameters and characteristics of this method and its potential for evaluation of the changes of the autonomic nervous system activity in different physiological and pathological conditions such as exogenous hypoxia, physical exercise and sleep. The application of heart rate variability in daily clinical practice would be beneficial for the diagnostics, the outcome prognosis and the assessment of the effect of treatment in various diseases.
Article
Full-text available
BACKGROUND: Heart rate variability (HRV) is an objective physiological marker that may be useful for monitoring training status in athletes. However, research aiming to interpret daily HRV changes in female athletes is limited. The objectives of this study were (1) to assess daily HRV (i.e., log-transformed root mean square of successive R-R interval differences, lnRMSSD) trends both as a team and intra-individually in response to varying training load (TL) and (2) to determine relationships between lnRMSSD fluctuation (coefficient of variation, lnRMSSDcv) and psychometric and fitness parameters in collegiate female soccer players (n=10). METHODS: Ultra-short, Smartphone-derived lnRMSSD and psychometrics were evaluated daily throughout 2 consecutive weeks of high and low TL. After the training period, fitness parameters were assessed. RESULTS: When compared to baseline, reductions in lnRMSSD ranged from unclear to very likely moderate during the high TL week (effect size ± 90% confidence limits [ES ± 90% CL] = -0.21 ± 0.74 to -0.64 ± 0.78, respectively) while lnRMSSD reductions were unclear during the low TL week (ES ± 90% CL = -0.03 ± 0.73 to -0.35 ± 0.75, respectively). A large difference in TL between weeks was observed (ES ± 90% CL = 1.37 ± 0.80). Higher lnRMSSDcv was associated with greater perceived fatigue and lower fitness (r [upper and lower 90% CL] = -0.55 [-0.84, -0.003] large, -0.65 [-0.89, -0.15] large). CONCLUSIONS: Athletes with lower fitness or higher perceived fatigue demonstrated greater reductions in lnRMSSD throughout training. This information can be useful when interpreting individual lnRMSSD responses throughout training for managing player fatigue.
Article
Full-text available
ABSTRACT Monitoring individual responses throughout training may provide insight to coaches regarding how athletes are coping to the current program. It is unclear if the evolution of heart rate variability (HRV) throughout training in team sport athletes can be useful in providing early indications of individual adaptation. The current study evaluated relationships between changes in resting cardiac-autonomic markers derived from a novel smartphone device within the first 3 weeks of a 5-week conditioning program and the eventual change in intermittent running performance at week 5 among 12 collegiate female soccer players. Change variables from week 1 to week 3 of the weekly mean and weekly coefficient of variation (CV) for resting heart rate (∆RHRmean and ∆RHRcv, respectively) and log transformed root mean square of successive R-R intervals multiplied by 20 (∆Ln rMSSDmean and ∆Ln rMSSDcv, respectively) were compared to changes in Yo-Yo Intermittent Recovery Test Level 1 performance (∆Yo-Yo). A very large and significant correlation was found between ∆Yo-Yo and ∆Ln rMSSDcv (r = -0.74; p = <0.01) and a large, non-significant correlation was found with ∆Ln rMSSDmean (r = 0.50; p = 0.096). This study suggests that a decrease in Ln rMSSDcv within the first 3 weeks of training is a favorable response, indicative of positive adaptation. Collecting daily HRV data with a smartphone application utilizing ultra-short HRV measures appears useful for athlete monitoring. KEY WORDS: HRV, Vagal, Parasympathetic, Yo-Yo, Women, Intermittent Running
Article
Full-text available
The aim of this study was to test the possibility of the ultra-short-term lnRMSSD (measured in 1-min post-1-min stabilization period) to detect training induced adaptations in futsal players. Twenty-four elite futsal players underwent HRV assessments pre- and post-three or four weeks preseason training. From the 10-min HRV recording period, lnRMSSD was analyzed in the following time segments: 1) from 0-5 min (i.e., stabilization period); 2) from 0-1 min; 1-2 min; 2-3 min; 3-4 min; 4-5 min and; 3) from 5-10 min (i.e., criterion period). The lnRMSSD was almost certainly higher (100/00/00) using the magnitude-based inference in all periods at the post- moment. The correlation between changes in ultra-short-term lnRMSSD (i.e., 0-1 min; 1-2 min; 2-3 min; 3-4 min; 4-5 min) and lnRMSSDCriterion ranged between 0.45 - 0.75, with the highest value (ρ = 0.75; 90% CI: 0.55 – 0.85) found between ultra-short-term lnRMDSSD at 1-2 min and lnRMSSDCriterion. In conclusion, lnRMSSD determined in a short period of 1-min is sensitive to training induced changes in futsal players (based on the very large correlation to the criterion measure), and can be used to track cardiac autonomic adaptations.
Article
Full-text available
This study evaluated the 7-day mean and CV of supine and standing ultra-short log transformed root mean square of successive R-R intervals multiplied by 20 (lnRMSSDx20) obtained with a smartphone application (app) in response to varying weekly training load (TL). Additionally, we aimed to determine if these values could be accurately assessed in as few as 5 or 3 days per week. 9 females from a collegiate soccer team performed daily HRV measures with an app in supine and standing positions over 3 weeks of moderate, high and low TL. The mean and CV over 7, 5, and 3 days were compared within and between each week. The 5 and 3-day measures within each week provided very good to near perfect intraclass correlations (ICCs ranging from 0.74 - 0.99) with typical errors ranging from 0.64 - 5.65 when compared with the 7-day criteria. The 7, 5, and 3-day supine CV and the 7-day standing CV were moderately lower during the low load compared to the high load week (p values ranged from 0.003 - 0.045 and effect sizes ranged from 0.86 - 0.92), with no significant changes occurring in the other measures. This study supports the use of the mean and CV of lnRMSSD measured across at least 5 days for reflecting weekly values. The supine lnRMSSDx20 CV as measured across 7, 5, and 3 days was the most sensitive marker to the changes in TL within the 3-week period.
Article
Heart rate variability (HRV) is a biomarker used to reflect both healthy and pathological state(s). The effect of the menstrual cycle and menstrual cycle phases (follicular, luteal) on HRV remains unclear. Active eumenorrheic women free from exogenous hormones completed five consecutive weeks of daily, oral basal body temperature (BBT) and HRV measurements upon waking. Descriptive statistics were used to characterize shifts in the HRV measures: Standard deviation of NN intervals (SDNN), root mean square of successive difference (rMSSD), high (HF) and low frequency (LF) across the menstrual cycle and between phases. All HRV measures were assessed by medians (Mdn), median difference of consecutive days (Mdnâ†) and variance. Seven participants (M ± SD; age: 28.60 ± 8.40 year) completed the study with regular menstrual cycles (28.40 ± 2.30 days; ovulation day 14.57 ± 0.98 day). Median rMSSD displayed a nonlinear decrease across the menstrual cycle and plateau around the day of ovulation. A negative shift before ovulation in Mdnâ†, rMSSD, SDNN, and LF as well as peak on luteal phase Day 4 in rMSSD and SDNN was observed. Median variance increased in rMSSD (150.06 ms ² ) SDNN (271.12 ms ² ), and LF variance (0.001 sec ² /Hz) from follicular to luteal phase. Daily HRV associated with the parasympathetic nervous system was observed to decrease nonlinearly across the menstrual cycle.
Article
The adaptive responses of the cardiovascular system to altitude appear to be dominated by increased sympathetic neural activity. We investigated the combined roles of the sympathetic and parasympathetic nervous systems (SNS and PNS, respectively) in the early (days 4–5) and subsequent (days 11–12) phases of acclimatization on Pike's Peak, CO (4,300 m), by spectral analysis of heart rate variability. Male subjects were randomly assigned to groups receiving oral propranolol (240 mg/day; n = 6) or a matched placebo (n = 3). On ascent to altitude, the high-frequency, fractal, and total spectral powers were reduced in the placebo group during days 4–5 and 11–12. At altitude during days 4–5, all three placebo group subjects increased SNS and decreased PNS activities compared with at sea level, and during days 11–12 SNS decreased and PNS increased compared with days 4–5. Relative to the placebo group, propranolol caused lengthening of the R-R interval; increases in high-frequency power, total spectral power, and the PNS indicator; and a decrease in the SNS indicator. Total spectral power tended to decrease at altitude, but there were no effects of altitude on PNS and SNS indicators in the propranolol group. The data from the placebo and propranolol groups suggest that both the PNS and SNS are involved in the elevated heart rate during the early phase of altitude acclimatization. Changes in heart rate variability during days 11–12 at altitude must be considered in light of the possible reductions in sympathetic receptor number noted in previous studies.
Article
In brief: An elevated resting pulse rate is generally considered a marker of overtraining in endurance athletes who greatly increase their workout distance. This study supports that assumption, demonstrating increased morning heart rates in 12 men who ran twice their regular training mileage during a 500-km (312-mile) road race over 20 days. After the first week of running, morning pulse rates were slightly reduced, but thereafter they progressively increased, becoming 10 beats min−1 higher (p <.01) as the race ended. Blood pressure, oral temperature, body weight, sweat loss, and blood glucose, lactate, insulin, and Cortisol levels were not related to the increase in morning heart rate.
Article
The aim of this study was to investigate factors that can predict individual adaptation to high-volume or high-intensity endurance training. After the first 8-week preparation period, 37 recreational endurance runners were matched into the high-volume training group (HVT) and high-intensity training group (HIT). During the next 8-week training period, HVT increased their running training volume and HIT increased training intensity. Endurance performance characteristics, heart rate variability (HRV), and serum hormone concentrations were measured before and after the training periods. While HIT improved peak treadmill running speed (RSpeak ) 3.1 ± 2.8% (P < 0.001), no significant changes occurred in HVT (RSpeak : 0.5 ± 1.9%). However, large individual variation was found in the changes of RSpeak in both groups (HVT: -2.8 to 4.1%; HIT: 0-10.2%). A negative relationship was observed between baseline high-frequency power of HRV (HFPnight ) and the individual changes of RSpeak (r = -0.74, P = 0.006) in HVT and a positive relationship (r = 0.63, P = 0.039) in HIT. Individuals with lower HFP showed greater change of RSpeak in HVT, while individuals with higher HFP responded well in HIT. It is concluded that nocturnal HRV can be used to individualize endurance training in recreational runners. © 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.